docstrings and unit tests (583865c2) · Commits · 钟慕尧 / deepchem

deepchem/models/tensorflow_models/IRV.py

+14 −3

Original line number	Diff line number	Diff line
		@@ -79,11 +79,20 @@ class IRVLayer(Layer):


		class IRVRegularize(Layer):
		""" This Layer extracts the trainable weights in IRVLayer
		and return the their L2-norm
		""" Extracts the trainable weights in IRVLayer
		and return their L2-norm
		No in_layers is required, but should be built after target IRVLayer
		"""

		def __init__(self, IRVLayer, penalty=0.0, **kwargs):
		"""
		Parameters
		----------
		IRVLayer: IRVLayer
		Target layer for extracting weights and regularization
		penalty: float
		L2 Penalty strength
		"""
		self.IRVLayer = IRVLayer
		self.penalty = penalty
		super(IRVRegularize, self).__init__(**kwargs)
		@@ -100,7 +109,9 @@ class IRVRegularize(Layer):


		class Slice(Layer):
		""" Choose a slice of input given axis and order
		""" Choose a slice of input on the last axis given order,
		Suppose input x has two dimensions,
		output f(x) = x[:, slice_num:slice_num+1]
		"""

		def __init__(self, slice_num, axis=1, **kwargs):

deepchem/models/tensorgraph/layers.py

+9 −0

Original line number	Diff line number	Diff line
		@@ -1185,6 +1185,9 @@ class SoftMax(Layer):


		class Sigmoid(Layer):
		""" Compute the sigmoid of input: f(x) = sigmoid(x)
		Only one input is allowed, output will have the same shape as input
		"""

		def __init__(self, in_layers=None, **kwargs):
		super(Sigmoid, self).__init__(in_layers, **kwargs)
		@@ -1552,6 +1555,12 @@ class SoftMaxCrossEntropy(Layer):


		class SigmoidCrossEntropy(Layer):
		""" Compute the sigmoid cross entropy of inputs: [labels, logits]
		`labels` hold the binary labels(with no axis of n_classes),
		`logits` hold the log probabilities for positive class(label=1),
		`labels` and `logits` should have same shape and type.
		Output will have the same shape as `logits`
		"""

		def __init__(self, in_layers=None, **kwargs):
		super(SigmoidCrossEntropy, self).__init__(in_layers, **kwargs)

deepchem/models/tensorgraph/tests/test_layers.py

+59 −0

Original line number	Diff line number	Diff line
		@@ -40,6 +40,8 @@ from deepchem.models.tensorgraph.layers import ReduceSum
		from deepchem.models.tensorgraph.layers import Repeat
		from deepchem.models.tensorgraph.layers import Reshape
		from deepchem.models.tensorgraph.layers import SluiceLoss
		from deepchem.models.tensorgraph.layers import Sigmoid
		from deepchem.models.tensorgraph.layers import SigmoidCrossEntropy
		from deepchem.models.tensorgraph.layers import SoftMax
		from deepchem.models.tensorgraph.layers import SoftMaxCrossEntropy
		from deepchem.models.tensorgraph.layers import StopGradient
		@@ -52,6 +54,10 @@ from deepchem.models.tensorgraph.layers import VinaFreeEnergy
		from deepchem.models.tensorgraph.layers import WeightedError
		from deepchem.models.tensorgraph.layers import WeightedLinearCombo

		from deepchem.models.tensorflow_models.IRV import IRVLayer
		from deepchem.models.tensorflow_models.IRV import IRVRegularize
		from deepchem.models.tensorflow_models.IRV import Slice


		class TestLayers(test_util.TensorFlowTestCase):
		"""
		@@ -223,6 +229,17 @@ class TestLayers(test_util.TensorFlowTestCase):
		out_tensor = out_tensor.eval()
		assert out_tensor.shape == (batch_size,)

		def test_sigmoid(self):
		"""Test that Sigmoid can be invoked."""
		batch_size = 10
		n_features = 5
		in_tensor = np.random.rand(batch_size, n_features)
		with self.test_session() as sess:
		in_tensor = tf.convert_to_tensor(in_tensor, dtype=tf.float32)
		out_tensor = Sigmoid()(in_tensor)
		out_tensor = out_tensor.eval()
		assert out_tensor.shape == (batch_size, n_features)

		def test_softmax(self):
		"""Test that Softmax can be invoked."""
		batch_size = 10
		@@ -327,6 +344,19 @@ class TestLayers(test_util.TensorFlowTestCase):
		dists = dist_tensor.eval()
		assert dists.shape == (N_atoms, M_nbrs)

		def test_sigmoid_cross_entropy(self):
		"""Test that SigmoidCrossEntropy can be invoked."""
		batch_size = 10
		n_features = 5
		logit_tensor = np.random.rand(batch_size, n_features)
		label_tensor = np.random.randint(0, 2, (batch_size, n_features))
		with self.test_session() as sess:
		logit_tensor = tf.convert_to_tensor(logit_tensor, dtype=tf.float32)
		label_tensor = tf.convert_to_tensor(label_tensor, dtype=tf.float32)
		out_tensor = SigmoidCrossEntropy()(label_tensor, logit_tensor)
		out_tensor = out_tensor.eval()
		assert out_tensor.shape == (batch_size, n_features)

		def test_softmax_cross_entropy(self):
		"""Test that SoftMaxCrossEntropy can be invoked."""
		batch_size = 10
		@@ -805,3 +835,32 @@ class TestLayers(test_util.TensorFlowTestCase):
		vertex_props, adjs = vertex_props.eval(), adjs.eval()
		assert vertex_props.shape == (10, 6, 50)
		assert adjs.shape == (10, 6, 5, 6)

		def test_slice(self):
		"""Test that Slice can be invoked."""
		batch_size = 10
		n_features = 5
		test_tensor_input = np.random.rand(batch_size, n_features)
		with self.test_session() as sess:
		test_tensor = tf.convert_to_tensor(test_tensor_input, dtype=tf.float32)
		out_tensor = Slice(1)(test_tensor)
		out_tensor = out_tensor.eval()
		assert np.allclose(out_tensor, test_tensor_input[:, 1:2])

		def test_IRV(self):
		"""Test that IRVLayer and IRVRegularize can be invoked."""
		batch_size = 10
		n_tasks = 5
		K = 10
		n_features = 2 * K * n_tasks
		test_tensor_input = np.random.rand(batch_size, n_features)
		with self.test_session() as sess:
		test_tensor = tf.convert_to_tensor(test_tensor_input, dtype=tf.float32)
		irv_layer = IRVLayer(n_tasks, K)
		irv_layer.create_tensor(in_layers=[test_tensor])
		out_tensor = irv_layer.out_tensor
		sess.run(tf.global_variables_initializer())
		out_tensor = out_tensor.eval()
		assert out_tensor.shape == (batch_size, n_tasks)
		irv_reg = IRVRegularize(irv_layer, 1.)()
		assert irv_reg.eval() >= 0

Admin message